Road Extraction on SpaceNet
0.6142mIoUFCCLIP
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| FCCLIPBackbone=ConvNeXt-L, Type=OVS2026.04 | 0.6142 | — | — | 79.95 | |
| EOV-SEGBackbone=ConvNeXt-L, Type=OVS2026.04 | 0.6052 | — | — | 78.59 | |
| GSNETBackbone=ViT-L, Type=OVRSIS2026.04 | 0.6018 | — | — | 81.04 | |
| Pi-SegBackbone=ViT-L, Type=OVRSIS2026.04 | 0.6017 | — | — | 81.64 | |
| FCCLIPBackbone=ConvNeXt-B, Type=OVS2026.04 | 0.5992 | — | — | 79.98 | |
| Pi-SegBackbone=ViT-B, Type=OVRSIS2026.04 | 0.594 | — | — | 74.66 | |
| OVRSBackbone=ViT-B, Type=OVRSIS2026.04 | 0.5781 | — | — | 70.49 | |
| CAT-SEGBackbone=ViT-B, Type=OVS2026.04 | 0.5743 | — | — | 78.32 | |
| CAT-SEGBackbone=ViT-L, Type=OVS2026.04 | 0.5729 | — | — | 81.73 | |
| SEDBackbone=ConvNeXt-B, Type=OVS2026.04 | 0.5707 | — | — | 72.44 | |
| SEDBackbone=ConvNeXt-L, Type=OVS2026.04 | 0.5675 | — | — | 73.52 | |
| EOV-SEGBackbone=ConvNeXt-B, Type=OVS2026.04 | 0.5648 | — | — | 76.22 | |
| SANBackbone=ViT-B, Type=OVS2026.04 | 0.5581 | — | — | 68.91 | |
| OVRSBackbone=ViT-L, Type=OVRSIS2026.04 | 0.553 | — | — | 80.86 | |
| SANBackbone=ViT-L, Type=OVS2026.04 | 0.5507 | — | — | 77.6 | |
| SCANBackbone=ViT-L, Type=OVS2026.04 | 0.548 | — | — | 74.53 | |
| RSKT-SEGBackbone=ViT-B, Type=OVRSIS2026.04 | 0.5336 | — | — | 71.95 | |
| GSNETBackbone=ViT-B, Type=OVRSIS2026.04 | 0.5273 | — | — | 77.95 | |
| SCANBackbone=ViT-B, Type=OVS2026.04 | 0.5236 | — | — | 69.8 | |
| RSKT-SEGBackbone=ViT-L, Type=OVRSIS2026.04 | 0.4887 | — | — | 74.7 | |
| SegEarth-OVResolution=896 x 8962024.10 | 0.291 | — | — | — | |
| SegEarth-OVResolution=448 x 4482024.10 | 0.238 | — | — | — | |
| SegEarth-OVBackbone=ViT-B, Type=OVRSIS (TF)2026.04 | 0.238 | — | — | — | |
| MaskCLIPResolution=448 x 4482024.10 | 0.208 | — | — | — | |
| MaskCLIPBackbone=ViT-B, Type=OVS (TF)2026.04 | 0.208 | — | — | — | |
| ClearCLIPResolution=448 x 4482024.10 | 0.163 | — | — | — | |
| ClearCLIPBackbone=ViT-B, Type=OVS (TF)2026.04 | 0.163 | — | — | — | |
| SCLIPResolution=448 x 4482024.10 | 0.149 | — | — | — | |
| SCLIPBackbone=ViT-B, Type=OVS (TF)2026.04 | 0.149 | — | — | — | |
| GEMResolution=448 x 4482024.10 | 0.119 | — | — | — | |
| GEMBackbone=ViT-B, Type=OVS (TF)2026.04 | 0.119 | — | — | — | |
| CLIPResolution=448 x 4482024.10 | 0.071 | — | — | — | |
| CLIPBackbone=ViT-B, Type=OVS (TF)2026.04 | 0.071 | — | — | — | |
| ClearCLIPTraining/Evaluation Protocol=Zero-shot2026.02 | — | 0.163 | — | — | |
| ClearCLIP2026.03 | — | — | 17.02 | — | |
| CLIPTraining/Evaluation Protocol=Zero-shot2026.02 | — | 0.071 | — | — | |
| CLIP2026.03 | — | — | 3.2 | — | |
| CLIP-SAMTraining Protocol=Zero-shot, Input Resolution=896x8962026.02 | — | 0.295 | — | — | |
| ConInfer2026.03 | — | — | 24.02 | — | |
| GEMTraining/Evaluation Protocol=Zero-shot2026.02 | — | 0.119 | — | — | |
| GEM2026.03 | — | — | 17.92 | — | |
| MaskCLIPTraining/Evaluation Protocol=Zero-shot2026.02 | — | 0.208 | — | — | |
| MaskCLIP2026.03 | — | — | 17.39 | — | |
| MaskCLIP*configuration=baseline2026.03 | — | — | 19.83 | — | |
| OursTraining/Evaluation Protocol=Zero-shot2026.02 | — | 0.261 | — | — | |
| SCLIPTraining/Evaluation Protocol=Zero-shot2026.02 | — | 0.149 | — | — | |
| SCLIP2026.03 | — | — | 8.13 | — | |
| SegEarth-OVTraining Protocol=Trained on remote sensing data, Input Resolution=896x8962026.02 | — | 0.291 | — | — | |
| SegEarth-OVTraining/Evaluation Protocol=Trained on remote sensing data2026.02 | — | 0.238 | — | — | |
| SegEarth-OV2026.03 | — | — | 20.85 | — |